I'm working with QC 10 and QTP 11. We have a suite of regression tests (around 800) that takes about 11 hours to execute. However, most of the time the full set of tests fails to execute. Typically around 8 hours into the test execution QC will start reporting errors about being unable to download tests. All tests thereafter will fail with the same error.

I saw some posts on the QTP forum that suggest that QC is unreliable when executing large test sets and some people have gone so far as to create their own test driver utilities as a replacement for QC. Anyone here have any comments on your own experiences with QC, QTP and large test sets? Does it work for you or have others seen similar problems?